It's not like meters to feet, the only way to measure performance is relative to the people you play.
Converting Glicko to Elo
Sort:
Jun 10, 2016
0
#3
I understand but there must be a way to compare it... I suppose a given Gicko score should be more o less equivalent to an ELO one...
It is not Elo and Glicko that you should be comparing, but rather the pools of players who are competing against each other on a regular basis, eg. the players who participate in FIDE OTB tournaments vs. the online players here. There are some differences between Elo and Glicko as systems, eg. Glicko takes into account the greater uncertainty of the ratings of a player who has played few games, but the main difference in the ratings stems from the different pools.
NM Matt Jensen compared ratings for chess.com, USCF and FIDE in this blog post. You'll noticed that under 1700, it is relatively easy to get a good FIDE rating while over 1700 it becomes more difficult compared to USCF.
Jun 11, 2016
0
#5
Thanks wayne,
I understand how both work. But I assume that in equals conditions (for example if a player plays a really big number of games or if he plays against the same opponents but ranked by both systems) both systems should be comparable, since they are evaluating a measurable capability that does not really change very quick (assuming a statistically representative number of games).
I understand how both work. But I assume that in equals conditions (for example if a player plays a really big number of games or if he plays against the same opponents but ranked by both systems) both systems should be comparable, since they are evaluating a measurable capability that does not really change very quick (assuming a statistically representative number of games).
When you say your "Glicko score" is around 1250, which federation or website did this calculation? Also which federation or website will be doing the calculation of your Elo rating? Are Matt Jensen's comparisons what you are looking for?
Jun 11, 2016
0
#7
sebas1789 wrote:
If he plays against the same opponents but ranked by both systems
If he plays against the same opponents but ranked by both systems
Interesting hypothetical. I'd guess the two ratings would be extremely close.
Most people who ask aren't asking about Elo to Glicko, they're asking about e.g. chess.com to FIDE (chess.com uses Glicko and FIDE uses Elo). If you're actually asking Elo to Glicko that's an interesting thought.
I think the USCF (which currently uses a system designed by Glickman) treats Elo ratings coming from FIDE or CFC as if they were similar to a USCF rating, so they must think the ratings are similar. In reality though, an individual's ratings coming from different federations tends to vary, sometimes quite a bit (GM Bator Sambuev was at one point rated 2753 CFC while his peak FIDE rating is around 2546).
There are many reasons why ratings vary. A player may play more games for one federation, or may play against weaker players at home, and stronger players internationally. Elo vs. Glicko or even different flavours of Elo or Glicko. Some federations or websites use ratings from other federations, and some don't. Some assign provisional ratings, while others assign you a rating right away. Some inject points into the rating pool to compensate for perceived inflation or deflation or in order to bring their ratings in line with another federation. Some penalize players whose ratings have drifted out of whack with the main pool because of their isolation. Not to mention that the players themselves probably play at different levels under different time controls or online vs. otb.
A rating is just a summary of your performance against the players you have faced in that federation.
Jun 12, 2016
0
#10
I asked just out of curiosity since I wanted to know how my score in websites like chess.com or chesscube could be extrapolated to the 'real world'
In Matt Jensen's stats, a 1200 chess.com blitz rating seemed to correspond with 1432 FIDE.
wayne_thomas wrote:
In Matt Jensen's stats, a 1200 chess.com blitz rating seemed to correspond with 1432 FIDE.
Nah. That is deffinetely not a general correlation. If that is what you are getting?
Feb 11, 2017
0
#13
BTW, it is Elo not ELO. Named for its originator, Arpad Elo. The USCF uses a modified version of Elo's system derived by Mark Glickman he calls the "Glicko" system. These are estimates of estimates, though, since all ratings are estimations of strength based on a specific pool of players. The pool of players in FIDE are not the same pool of players in the USCF. The pool of players on chess.com are not the same as the pool of players on ICC.
You can find the relative ratings using the formulas given at the bottom link, but here they are to save time:
USCF = 720 + 0.625*FIDE if FIDE < 2000 USCF = -350 + 1.16*FIDE if FIDE >= 2000
Two formulas depending on whether the FIDE rating is less than or greater than (or equal to) 2000. Of course you can solve for FIDE to get the reverse.
http://www.glicko.net/ratings/report08.txt
My otb is 1174 (rapid)(standard is 1173). Chess.com is around 1600 (rapid) and around 2000 on lichess (rapid). No idea how these translate.
Aug 25, 2020
0
#16
After playing many games here, your expected FIDE rating would be around 200 or 300 points less than chess.com rating.
Aug 25, 2020
0
#17
hanifjaan wrote:
My Lichess Bltz rating is 1800 but is 1500 on chess.com
It's normal, since you start in chess.com with 1200 rating while in Lichess it's 1500
Sep 10, 2023
0
#18
I played 1000 games on chess.com as a first time chess player at age 44. I am 700 rating. Anyone know with the information given what that might correlate to in other rating systems like elo. (Ballpark)
A rapid rating of 1330 on chess.com corresponds to a FIDE rating of around 1260 according to Matt Jensen's charts a few years ago.
Sep 12, 2023
0
#20
Mouselip wrote:
BTW, it is Elo not ELO. Named for its originator, Arpad Elo. The USCF uses a modified version of Elo's system derived by Mark Glickman he calls the "Glicko" system. These are estimates of estimates, though, since all ratings are estimations of strength based on a specific pool of players. The pool of players in FIDE are not the same pool of players in the USCF. The pool of players on chess.com are not the same as the pool of players on ICC.
You can find the relative ratings using the formulas given at the bottom link, but here they are to save time:
Two formulas depending on whether the FIDE rating is less than or greater than (or equal to) 2000. Of course you can solve for FIDE to get the reverse.
Your link from Glickman's files is pretty old. It appears to be from 2007, which is about when Chess.com appeared online with its current ownership.
It seems likely that there are more recent comparisons of the Elo ratings systems used by FIDE, the USCF and Chess.com. (The USCF and Chess.com use a modified Elo system developed by Professor Glickman.)
My Glicko score is around 1250, and I'm curious about how it correlates with ELO system. I mean what do you think would be my score according to ELO?
Thanks!!